Abstract. 1. Non-random patterns of species co-occurrences among habitats and the underlying mechanisms responsible for these patterns are fundamental to ecological research. Fewer co-occurrences than expected from a random null model are often interpreted as indicating that communities are structured by negative biotic interactions (ecological checkerboards) or by abiotic factors (habitat checkerboards). The difficulty has often been determining the dominant causative mechanism(s) behind these patterns.
2. In the current study, we examine patterns of insect co-occurrence in 23 low-gradient, sandy-bottom coastal streams of the Gulf of Mexico. Focusing our attention on five commonly collected orders in these sites, we ask two fundamental questions about the community: (i) Are patterns of insect co-occurrence non-random? (ii) If these patterns are non-random, are they consistent with the expectations of a community structured largely by abiotic conditions, biotic conditions (e.g. species interaction) or both?
3. Using a series of null models, we found a strong pattern of negative co-occurrence (segregation) for both genera and species when all stream sites were considered. When we partially controlled for the influence of abiotic conditions by examining a subset of streams with similar abiotic conditions, both species and genera showed a very significant pattern of aggregation. These results therefore suggest that abiotic conditions act as a strong filter, influencing the patterns of co-occurrence for both genera and species.
4. Comparisons of co-occurrence patterns between the species-level and genus-level analyses suggested similar degrees of segregation. Thus, no compelling evidence of competition influencing insect distributions among streams was found.